1986 Timoshenko Medal Acceptance Speech by George R. Irwin
Comments on Discovery and Invention
Text of a talk delivered at the Applied Mechanics Dinner of the 1986 Winter Annual Meeting of ASME in Anaheim, California.
With regard to the topic of these comments, I was told that a title of some kind was mandatory so I gave a title which seemed reasonably impressive. Upon reflection, I have little to offer in the way of comments which are correspondingly impressive. It occurred to me that the forward motion of a crack in a structural material might be of some interest as a descriptive model. A 1950 technical paper, by Kies, Sullivan, and Irwin, reported that progressive fracturing usually occurs by the development and joining of advance separations and that these local behaviors tend to be rather abrupt. By use of motivation as a driving force and by substituting "advance ideas" for "advance separations," a plausible descriptive model of forward technological progress seemed possible. Of course, details related to the development and joining of advance ideas would be needed. These would include motivation, opportunity, guidance, and information exchange. In his 1985 Timoshenko Medal comments, Sternberg noted certain research management features which are not helpful. The conditions one likes for best progress certainly include benign methods of research management. After additional reflection on these and other complexities related to innovative progress, I decided that the descriptive model I had thought to develop was unlikely to be useful. So my comments will have a different nature. They will be memories and historic fragments related to my topic and they will be restricted to the strength of materials field.
A 1955 "History of Strength of Materials" by Timoshenko contained comments on investigations by da Vinci, Galileo, Marriotte, and others. This book contained a rough copy of one of da Vinci's drawings which was of special interest to me because it portrayed a fracture strength testing machine. Leonardo da Vinci was born in 1452. Five hundred years later, in 1952, a large international meeting was held devoted to da Vinci. One permanent result was a splendid large book from Reynal containing authentic copies of da Vinci paintings and sketch book drawings. A description of his career was included. From this I learned that, as a young man, he was removed from a school for clerks at the request of the school master. A position was then found for him in an art shop. His professional development from that point was, as you know, remarkable. With regard to his interest in fracture strength, at that time the process of making iron wire was automated only to a limited degree. The wire was gripped by tongs and drawn through a die, segment by segment, by a man seated in a swing. He started each segment by pushing with his legs, then swung back to take another grip for the next segment. da Vinci must have been acquainted with this process and may have had doubts as to the reliability of the product. In one of his sketch books, along with other drawings from his inventive mind, one finds a relatively simple device for testing the fracture strength of iron wire. His comments, written at one side of this sketch, have the form of instructions to someone who is conducting the testing as follows:
"Observe what the weight was that broke the wire and in what part the wire broke. Then shorten the wire, at first by half, and see how much more weight it supports; and then make it one-quarter of its original length, and so on, making various lengths and noting the breaking weight and the place at which it breaks."
There is an inference here that short lengths of wire had larger fracture strength than long lengths. Given a population of flaws of variable severity, a result of this nature would be expected on an average basis. However, da Vinci's comments do not suggest a sue effect observation which is variable in nature. Furthermore, one can note that a positive increase of strength, with shortening of test length, would always be observed if the man conducting the test found it convenient to make each shorter length using the largest segment of wire provided by the previous test. As a testing report, da Vinci's comments are obviously incomplete. Understandably, Timoshenko found them to be confusing.
During the 1960s, relatively rapid progress was made toward structural use of fiber type composites This was accompanied by a renewal of interest in the effect of specimen length on the strength of fibers. The results consistently showed specimen length size effects, and it was clear that the strength of fibers depended very much upon flaws of variable size introduced by process handling. Optimum composite strengths were obtained when the fiber-to-matrix adhesion provided load transfer across flawed regions so that the effective fiber strength had the high value associated with very short fiber lengths. Despite the curious nature of his comments, the great Leonardo may have understood the influences of flaws upon strength of iron wire rather well, but we do not credit him with the discovery of statistical aspects of fracture sue effects. He perceived a problem, devised a useful test method, and obtained the answers he needed for engineering purposes.
Credit for a technological discovery is rarely given unless the originator communicates what was discovered with enough clarity and impact to attract useful applications of the results. I attended the 1946 International Meeting on Applied Mechanics in Paris. My paper discussed the decelerating force in a projectile during plate penetration. The only one there who expressed interest in this topic was Schardin. At the Berlin Hochschule he was a colleague of Cranz whose interest in military connected research was well known. The papers of major and general interest were given by G. I. Taylor and by Von Karman. Taylor discussed plastic wave propagation during impact of a cylinder with a plate. Von Karman discussed plastic wave propagation from suddenly applied tension in a wire. Their research was done quite independently and they arrived at similar expressions for plastic wave behavior in analytical form. These and subsequent related investigations have provided much of our information on plastic flow resistance of metals at high strain rates. Neither Taylor nor Von Karman seemed aware of an earlier paper by Donnell, where, in an appendix section, analytical expressions for plastic wave propagation were already available. You can think as you please regarding discovery credit. In regard to usefulness, communication clarity and impact do have substantial value.
Going back again to olden times, I have admired certain discoveries made by a French engineer, Marriotte. Marriotte was in charge of parts of the engineering work during construction of the well known palace and adjoining park at Versaille during the 1650 to 1680 period. In particular he was to provide fountains which had grandeur, that is, they propelled water to a large height, and that created a pressure vessel problem. Marriotte solved this problem in da Vinci fashion by proof testing. Pressure increases were supplied to cylindrical test vessels by extending lengths of stand-pipe upward. Possibly he selected a convenient hillside location. He reported his results to the French Academy of Science and they appeared in the proceedings of that Academy at about the same date as was selected by Hooke to publish the solution to his "Hooke's Law" anagram. Marriotte observed the fractional change of vessel circumference with increase of pressure and found a direct proportionality between them. He also reported that each vessel appeared to break when the fractional change of circumference had a fixed critical value. When I picture myself as one of his assistants, measuring circumferential strain, possibly with a tape around the vessel, I view his independent discovery of Hookers Law with much sympathy. I assume that this discovery permitted the last increments of pressure to be applied without need for direct stretch measurements. His suggestion of a critical strain condition for fracture inaugurated various methods of characterizing strength with stress-strain relationships.
Some old ideas, like old soldiers, fade away and others endure. Usually they acquire different costumes. Recent papers on dynamic fracturing by Freund and by Achenbach use Marriotte's fracture criterion in a special form. The key element related to survival of a concept is usefulness. Pushing old ideas into new applications is often appropriate and helpful. In 1950, Green and Sneddon published a solution for the stress-strain pattern around a flat elliptical crack subjected to normal tension. The authors said that the analysis method, which they applied to their problem, could be found in either of two papers published about 50 years earlier. One dealt with liquid flow past a flat-elliptical plate normal to the stream. The other pertained to the gravitational field of a similar flat- elliptical object. With the aid of several simplifying assumptions, the solution provided by Green and Sneddon was quite helpful in estimating the relative danger of part-through surface cracks in a plate. Linear-elastic fracture mechanics developed into a useful engineering tool during the twenty years following World War 11. It is noteworthy that the analysis methods needed for this development were all available either before or very early in that time period. Obviously scholarship and use of available knowledge should continue to assist technological advancement as it has in the past. The use of computer methods to provide numerical illustration of analysis concepts will help us move ahead more rapidly. An illustration was furnished by the rapid development of a new branch of fracture mechanics termed elastic-plastic fracture mechanics. Starting from technical papers by Rice and by Hutchinson, this development became of practical value in less than ten years.
Science began with looking and description, along with suitable methods of communicating that which was observed. We haven't exhausted the value of descriptive science in the field of strength of materials. However, the observed patterns of behavior, when sufficiently clear, lead frequently to analytical models and these are of basic importance. They help us understand the observed behaviors and they permit sensible planning in engineering applications.
I have only a few closing comments. I read through the list of 1986 ASME Award winners. Assuming as typical, those who were well known to me, we have a rich supply of engineers deserving of recognition for the useful nature of their work. The society members who gave their own time to recommendation or selection of Award winners also deserve praise. In 1946, I received an award from the Navy Department. They evidently thought I had done useful things. The idea that I could do things which were useful enough to deserve notice had a significant result. It changed me from being a physicist of some kind into becoming some kind of an engineer. In the various fields of engineering and especially for research activities, motivation is an important driving force and usefulness is an important goal.